Different stepsize for different parts of Modules / different mechanisms

[jnsbck]

This came up during a meeting: The idea was that for specific mechanisms, i.e. leak gates or other slow transient currents or for different parts of a module (a subset of cells with only leak currents for example) one could speed up computation, if one could set a larger dt for these than for others. I am not sure how one would implement this though. Opening this issue to keep this idea around.

[michaeldeistler]

We don't have to implement this for v1.0, but we should at least think about the API and how it will affect the Channel.